connect by 用于存在父子,祖孙,上下级等层级关系的数据表进行层级查询。
语法格式:
{ CONNECT BY [ NOCYCLE ] condition [AND condition]... [ START WITH condition ]
| START WITH condition CONNECT BY [ NOCYCLE ] condition [AND condition]...
}
特殊词讲解:
start with: 指定起始节点的条件
connect by: 指定父子行的条件关系
prior: 查询父行的限定符,格式: prior column1 = column2 or column1 = prior column2 and ... ,
nocycle: 若数据表中存在循环行,那么不添加此关键字会报错,添加关键字后,便不会报错,但循环的两行只会显示其中的第一条
循环行: 该行只有一个子行,而且子行又是该行的祖先行
connect_by_iscycle: 前置条件:在使用了nocycle之后才能使用此关键字,用于表示是否是循环行,0表示否,1 表示是
connect_by_isleaf: 是否是叶子节点,0表示否,1 表示是
level: level伪列,表示层级,值越小层级越高,level=1为层级最高节点
-- 创建表
create table employee(
emp_id number(18),
lead_id number(18),
emp_name varchar2(200),
salary number(10,2),
dept_no varchar2(8)
); -- 添加数据
insert into employee values('',0,'king','1000000.00','');
insert into employee values('',1,'jack','50500.00','');
insert into employee values('',1,'arise','60000.00','');
insert into employee values('',2,'scott','30000.00','');
insert into employee values('',2,'tiger','25000.00','');
insert into employee values('',3,'wudde','23000.00','');
insert into employee values('',3,'joker','21000.00','');
commit;
(1) 查询以lead_id为0开始的节点的所有直属节点
select emp_id,lead_id,emp_name,prior emp_name as lead_name,salary
from employee
start with lead_id=0
connect by prior emp_id = lead_id -- 等同于 select emp_id,lead_id,emp_name,prior emp_name as lead_name,salary
from employee
start with emp_id=1
connect by prior emp_id = lead_id
(2) 以emp_id为6的所有祖先节点
select emp_id,lead_id,emp_name,salary
from employee
start with emp_id=6
connect by prior lead_id=emp_id;
(3) 查询一个节点的叔叔伯父节点
--查看emp_id为6的节点的叔叔伯父节点
with temp as (
select employee.*,
prior emp_name,
level le
from employee
start with lead_id = 0
connect by lead_id=prior emp_id
)
select *
from temp t
left join temp tt
on tt.emp_id=6 --此处需要限定
where t.le = (tt.le-1)
and t.emp_id not in (tt.lead_id)
(4) 查询族兄
--查看employee id是6的节点的族兄节点
with temp as (
select employee.*,
prior emp_name,
level le
from employee
start with lead_id=0
connect by lead_id= prior emp_id
) select t.*
from temp t
left outer join temp tt
on tt.emp_id=6 --此处需要条件限制
where t.le=tt.le
and t.emp_id<>6 --此处需要条件限制
(5) level伪列的使用,格式化层级
select lpad(' ',level*2,' ')||emp_name as name,emp_id,lead_id,salary,level
from employee
start with lead_id=0
connect by prior emp_id=lead_id
level数值越低级别越高
(6) connect_by_root 查找根节点
select connect_by_root emp_name,emp_name,lead_id,salary
from employe
start with lead_id=1
connect by prior emp_id = lead_id;
注意: connect_by_root关键字后面跟着字段,表示根节点对应记录的某一字段的值,
如 connect_by_root emp_name表示根节点的员工名,connect_by_root salary表示根节点的工资
(7) 标注循环行
-- 插入一条数据,与另一条emp_id=7的数据组成循环行
insert into employee values('',7,'joker_cycle','21000.00','');
commit; -- connect_by_iscycle("CYCLE"), connect by nocycle
select emp_id,emp_name,lead_id,salary,connect_by_iscycle as cycle
from employee
start with lead_id=0
connect by nocycle prior emp_id = lead_id;
最后一行与新追加的一行是循环关系,因此connect_by_iscycle列显示值为1,但结果集只显示循环的第一条,与之循环的另外一条(新追加的一条)是不显示的
connect by 后面的nocycle关键字是防止出现父子关系循环的,如果表中出现父子关系循环且没有使用该关键字,会报如下错误:
--ORA-01436: 用户数据中的 CONNECT BY 循环
(8) connect_by_isleaf 是否是叶子节点
select emp_id,emp_name,lead_id,salary,connect_by_isleaf
from employee
start with lead_id=0
connect by nocycle prior emp_id=lead_id;
叶节点指的是没有子节点的节点,那些是既是父节点又是子节点的节点不属于叶节点